| noonedeadpunk | mornings | 10:24 |
|---|---|---|
| noonedeadpunk | I'm still trying to deploy damn compute node lol :D | 10:26 |
| jrosser | morning! | 10:27 |
| jrosser | i'm still fiddling with the log collection stuff which i thought would be pretty quick /o\ | 10:27 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Refactor log collection from lxc containers https://review.opendev.org/c/openstack/openstack-ansible/+/877382 | 10:42 |
| noonedeadpunk | ThiagoCMC: hey! regarding rally, can you try running `os-rally-install.yml -e venv_wheel_build_enable=true`? | 10:57 |
| noonedeadpunk | Also, the host where you're trying to install rally is running ubuntu 20.04, right? | 10:57 |
| opendevreview | Merged openstack/openstack-ansible-os_ironic stable/zed: Update ironic documentation https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/877393 | 11:02 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible master: Remove nova-scheduler from ironic compute container_skel https://review.opendev.org/c/openstack/openstack-ansible/+/877467 | 12:00 |
| hamidlotfi | Hi there, | 12:32 |
| hamidlotfi | I wanted to scale my OSA environment with your guide published at `https://docs.openstack.org/openstack-ansible/latest/admin/scale-environment.html` | 12:32 |
| hamidlotfi | But the new node and all containers contained in it do not add to /etc/hosts file of other nodes. | 12:32 |
| hamidlotfi | for example, I added a node called infra04.but infra04 does not pinging with itself name from infra01 or infra02 or infra03 . | 12:35 |
| noonedeadpunk | hamidlotfi: Well, if you was limiting all deployment to infra04 then there's no way how ansible playbook could run against infra01-03 | 12:44 |
| noonedeadpunk | So I'd run `openstack-ansible playbooks/openstack-hosts-setup.yml -e openstack_host_group=all --tags openstack_hosts-config | 12:47 |
| hamidlotfi | noonedeadpunk: because in that document exactly writed `To prepare new hosts and deploy containers on them run setup-hosts.yml playbook with the limit argument` | 12:47 |
| noonedeadpunk | Yeah, we might miss that step in the document | 12:48 |
| hamidlotfi | please explain it here. | 12:48 |
| noonedeadpunk | Eentually I think we miss it for compute deployment as well.... | 12:49 |
| noonedeadpunk | Well, as I said, in order to update hosts file on infra01, you need to run an openstack_hosts role against this host | 12:50 |
| noonedeadpunk | And if you're limiting to infra04 only - play won't touch anything except infra04 | 12:50 |
| noonedeadpunk | Can you report a bug regarding this documentation issue so we won't forget to fix it? | 12:51 |
| opendevreview | Merged openstack/openstack-ansible-haproxy_server stable/xena: Use let's encrypt standalone flag only for http-01 https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/874621 | 13:18 |
| noonedeadpunk | or propose fix to it :-) | 13:25 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_ironic master: Install socat and configure ipmtool-socat console interface https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/877479 | 13:52 |
| damiandabrowski | jrosser: regarding https://review.opendev.org/c/openstack/openstack-ansible/+/876639 | 14:08 |
| damiandabrowski | I'm not sure how much we should focus on saving CI resources, but do you think it would be worth to just add "stepca" scenario to two already existing jobs(one lxc and one metal) instead of defining two completely new jobs? | 14:08 |
| dokeeffe85 | Hi all, it's been a while :) quick question for anyone that may have an answer. I could create a flat network that had access to the internet fine and as many private networks as I wanted in my last test deploy and it all worked fine. Now my question is, using ovs can I create multiple "external" networks or would I handle that by using routers? The simplest way possible would suit, if it can be done that is!! Thanks in advance as always | 14:15 |
| noonedeadpunk | damiandabrowski: I think we also want to test our pki scenario as well. So adding stepci to existing jobs will prevent us from checking another path | 14:22 |
| jrosser | i wonder what the impact on CI resources is now its !tripleo | 14:22 |
| damiandabrowski | dokeeffe85: regardless of ml2 driver, you can create only one flat provider network for each physical network | 14:23 |
| noonedeadpunk | dokeeffe85: sorry, I'm not sure I got question... So you have multiple public interfaces/networks you want to add? | 14:23 |
| noonedeadpunk | yeah, but you can also create plenty of vlans external networks on the same flat if needed... | 14:24 |
| damiandabrowski | if you want to change that, I'd suggest leveraging vlans/vxlans/geneve etc. | 14:24 |
| noonedeadpunk | But I'm not sure I got what you really want to achieve to be frank | 14:25 |
| noonedeadpunk | As you can also add multiple subnets to the one external network | 14:25 |
| dokeeffe85 | Ok sorry, yeah what I wanted to do was create lets say 87.44.17, .18, .19, .20 all /24 and have 4 different networks with internet access. This is possible using vlan networks on one flat network?? Does it need to be set up on the original run of the playbooks or can I do it through openstack later?# | 14:41 |
| noonedeadpunk | you can do that anytime I would say. If you haven't configured vlan networks - you can do that afterwards but will need to re-run neutron playbook | 14:43 |
| noonedeadpunk | If vlan networks were configured already - the only thing you would need to do is to create networks with horizon/cli/api with correct type and vlan ids | 14:43 |
| opendevreview | Takashi Kajinami proposed openstack/ansible-config_template master: Remove TripleO jobs https://review.opendev.org/c/openstack/ansible-config_template/+/877486 | 14:44 |
| noonedeadpunk | `vlan networks were configured already` -> I mean mappings for ml2 plugin | 14:44 |
| opendevreview | Dmitriy Rabotyagov proposed openstack/openstack-ansible stable/xena: Bump OpenStack-Ansible Xena https://review.opendev.org/c/openstack/openstack-ansible/+/877488 | 14:47 |
| opendevreview | Takashi Kajinami proposed openstack/ansible-role-python_venv_build master: Remove TripleO jobs https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/877510 | 14:49 |
| opendevreview | Takashi Kajinami proposed openstack/ansible-role-python_venv_build stable/ussuri: Ussuri-only: Remove TripleO job https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/877511 | 14:54 |
| opendevreview | Takashi Kajinami proposed openstack/ansible-role-python_venv_build stable/victoria: Victoria-only: Remove TripleO job https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/877512 | 14:55 |
| admin1 | dokeeffe85, using vlan is the best, as you can then control/route the vlans | 15:46 |
| admin1 | and you can create as many networks as you want which are external | 15:46 |
| noonedeadpunk | Actually, no more then 4096 :D | 15:47 |
| noonedeadpunk | but yeah, quite a lot :D | 15:48 |
| admin1 | imaging 4000+ ext- networks :) | 15:48 |
| admin1 | imagine* | 15:48 |
| jrosser | just FYI i proved right now that `openstack_opendev_base_url: https://github.com` works a treat if your connectivity to opendev.org goes bad | 15:50 |
| dokeeffe85 | Thank you noonedead punk. Would I also need to do bonds in the netplan for each of those vlan networks, you wouldn't have a link to an example or similar would you please? | 16:00 |
| jrosser | dokeeffe85: the advantage of doing it with vlan networks is that you don't need to specify the actual vlan anywhere in the host networking config, it's a "trunk" port from your switch | 16:02 |
| jrosser | so to bring up a new vlan type external network you would permit that particular vlan tag on the switch ports | 16:02 |
| jrosser | then you'd create the vlan type provider network with segmentation_id being the vlan tag, and off you go | 16:03 |
| jrosser | ^ create the neuton network with the openstack cli | 16:04 |
| noonedeadpunk | admin1: well, they could be /28 or smth lol | 16:17 |
| ThiagoCMC | noonedeadpunk, yes, it's Ubuntu 20.04, I'll try your suggestion now. | 16:35 |
| ThiagoCMC | noonedeadpunk, worked! | 16:42 |
| noonedeadpunk | well, you can place that safely in user_variables if you want to) | 16:47 |
| ThiagoCMC | Ahh... Cool, thanks! | 16:51 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_swift master: DNM - test commit https://review.opendev.org/c/openstack/openstack-ansible-os_swift/+/877530 | 17:02 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_ironic master: DNM - test https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/877531 | 17:02 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_swift master: Fix role as a result of ANSIBLE_INJECT_FACT_VARS=false https://review.opendev.org/c/openstack/openstack-ansible-os_swift/+/877534 | 17:26 |
| jrosser | noonedeadpunk: we broke some stuff ^^ | 17:26 |
| jrosser | there will be more too | 17:26 |
| jrosser | like here https://github.com/openstack/openstack-ansible/blob/master/playbooks/common-tasks/dynamic-address-fact.yml#L19 | 17:27 |
| * noonedeadpunk not surprised | 17:27 | |
| jrosser | that makes it interesting to think we can just make ANSIBLE_INJECT_FACT_VARS=true again | 17:27 |
| jrosser | might have to craft the code specially to deal with both cases | 17:27 |
| noonedeadpunk | Wait, should not hostvars[inventory_hostname] work anyway? | 17:27 |
| noonedeadpunk | ah, it's fact | 17:28 |
| noonedeadpunk | ok, yes, sorry, disregard | 17:28 |
| jrosser | idk what that code is about from before, but very old | 17:28 |
| noonedeadpunk | damn, now we should think twice what is fact and what is var... | 17:29 |
| noonedeadpunk | as hostvars is no longer universal thing... | 17:30 |
| jrosser | again here :( https://github.com/openstack/openstack-ansible/blob/master/playbooks/common-tasks/dynamic-address-fact.yml#L30 | 17:31 |
| jrosser | oh no thats ok i think | 17:31 |
| opendevreview | Jonathan Rosser proposed openstack/openstack-ansible-os_ironic master: Update IPA version for 2023.1 release of Ironic https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/877538 | 17:44 |
| ThiagoCMC | It feels a bit weird that OSA uses CentOS 9 for `ironic_deploy_image_base_name` even in a 100% Ubuntu-based environment... :-P | 17:54 |
| jrosser | thats really down to the choice of the ironic team, they build it | 17:54 |
| ThiagoCMC | I understand it uses what's availalbe on opendev.org but, doesn't it bring side effects? I mean, a discovered machine with CentOS 9 can have different network interfaces name of a deployed Ubuntu, right? | 17:55 |
| ThiagoCMC | jrosser, sure... Just curious ^_^ | 17:55 |
| jrosser | well, IPA is a disk image you boot baremetal nodes with that then dd the actual OS you want onto the disk, and reboot | 17:55 |
| ThiagoCMC | I see... Sounds similar with MaaS | 17:55 |
| jrosser | its the provisioning agent for ironic | 17:56 |
| ThiagoCMC | Got it. I'm still very new with Ironic, didn't managed to make it work yet. lol | 17:56 |
| jrosser | i did a huge patch on the documentation recently which might help | 17:56 |
| ThiagoCMC | Nice! | 17:57 |
| ThiagoCMC | I'll give it a try again this month. Got new machines and I'm going for Ubuntu 22.04 and maybe Oracle Linux 9 too. | 17:57 |
| ThiagoCMC | Should I go with `stable/zed`, or maybe better stick with `master` (for playing with OSA + Ironic)? | 17:58 |
| opendevreview | Merged openstack/ansible-role-pki master: Allow to provide custom handler names https://review.opendev.org/c/openstack/ansible-role-pki/+/875757 | 18:06 |
| opendevreview | Merged openstack/openstack-ansible-haproxy_server master: Allow default_backend to be specified https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/876157 | 18:07 |
| jrosser | NeilHanlon: you wouldn't happen to know if there is an equivalent of this on RH-alikes? https://packages.ubuntu.com/kinetic/qemu-system | 18:18 |
| jrosser | qemu-system is a metapackage which pulls in * architecture qemu-system-* | 18:18 |
| jrosser | so if i want just x86 i can install qemu-system-x86 | 18:19 |
| NeilHanlon | I know it's in Fedora as of recent branches, but I don't think it made it into RHEL 9 | 18:21 |
| NeilHanlon | what the heck, i'll try to rebuild it lol | 18:21 |
| jrosser | hah well what i'm looking to do is be specific about the architectures that do get installed | 18:21 |
| jrosser | because the presence of qemu-system-<random-architecture> binary on a system of a different architecture has some really undesirable side effects | 18:22 |
| jrosser | ^ unless of course you are happy with totally random emulation of your VM on non-native architectures :) | 18:22 |
| NeilHanlon | ah, i see. i thought you were looking for like the qemu-static usermode emulation stuff | 18:23 |
| ThiagoCMC | Like with QEMU x86_64 VM on arm64 host? =P | 18:23 |
| NeilHanlon | in that case I think you can install `qemu-system-$arch-core` | 18:23 |
| jrosser | so i'd ideally like to be installing the arch-specific package based on the ansible_facts['architecture'], rather than just get everything | 18:23 |
| jrosser | ahha that'd do it - was just failing to find how they were named | 18:24 |
| NeilHanlon | ThiagoCMC: or worse, QEMU ppc64le on a ppc64le host 🙃 | 18:24 |
| jrosser | ThiagoCMC: that is *exactly* the sitaution i have here right now | 18:24 |
| jrosser | and it's been a total pain to make that not happen | 18:24 |
| NeilHanlon | yucky :( | 18:24 |
| ThiagoCMC | Cool! I think... lol | 18:25 |
| jrosser | if you have an existing cloud, like an x86 one | 18:25 |
| jrosser | and you have never had to put hw_architecture property on your images | 18:25 |
| jrosser | then as soon as you add an arm node all manner of wierd breaks out of qemu-system-x86 is installed on the arm node | 18:25 |
| ThiagoCMC | I always do that. | 18:25 |
| NeilHanlon | I'm checking to make sure the EL9 rpm is built in a way that does this.. it may have been a recent fedora change | 18:26 |
| ThiagoCMC | I mean, for example, setting properties 'cpu_arch: x86_64' | 18:27 |
| jrosser | i am not sure it is that simple | 18:27 |
| jrosser | really cpu_arch and not hw_architecture? | 18:29 |
| ThiagoCMC | https://docs.ansible.com/ansible/2.6/modules/os_image_module.html - It's `cpu_arch` | 18:30 |
| ThiagoCMC | At least with that Ansible module =P | 18:31 |
| jrosser | well tbh this sums it up unfortunatley https://www.mail-archive.com/openstack-dev@lists.openstack.org/msg109964.html | 18:34 |
| NeilHanlon | ookay.. on Rocky/RHEL 9, it looks like we only provide qemu-kvm from the base repositories. Where are we installing the qemu-system executables from? CentOS Virt sig? | 18:37 |
| jrosser | not sure tbh - i need to look | 18:38 |
| jrosser | our CI jobs make a log of whats installed and from where so that should be pretty straightforward to see | 18:38 |
| jrosser | NeilHanlon: https://zuul.opendev.org/t/openstack/build/e1ff3ed32af3419bbd019cf6d4125599/log/logs/redhat-rpm-list-installed-host-19-29-09.txt | 18:40 |
| jrosser | so in this case i think whatever we're doing we only get the x86 one https://zuul.opendev.org/t/openstack/build/e1ff3ed32af3419bbd019cf6d4125599/log/logs/redhat-rpm-list-installed-host-19-29-09.txt#546-562 | 18:41 |
| NeilHanlon | it appears that in theory, DNF should just install the right one based on the host architecture, and this issue shouldn't "exist" per se | 18:44 |
| * NeilHanlon should just rebuild latest qemu in COPR... | 18:45 | |
| NeilHanlon | btw jrosser: I assume that cloud-init thing has been working OK? I asked to upstream it into CentOS Stream, but am expecting it will take until after they release 9.2 and 8.8 in May before it gets merged in.. for now i'll keep the copr repo active | 18:51 |
| jrosser | NeilHanlon: yes it has been working ok | 19:05 |
| jrosser | NeilHanlon: you could check over this too (and the comments) to see if thats what you expect from DIB/Rocky | 19:07 |
| jrosser | https://review.opendev.org/c/openstack/openstack-ansible-os_ironic/+/867547 | 19:07 |
| jrosser | was quite an adventure to come up with a recipe that works for both x86 and aarch64 | 19:08 |
| jrosser | the whole DIB_CLOUD_INIT_DATASOURCES=Ec2 was unobvious, to say the least | 19:09 |
| NeilHanlon | definitely will take a look | 19:15 |
| opendevreview | Merged openstack/ansible-role-python_venv_build master: Remove TripleO jobs https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/877510 | 21:40 |
| opendevreview | Damian DÄ…browski proposed openstack/openstack-ansible master: Split haproxy horizon config into 'base' frontend and 'horizon' backend https://review.opendev.org/c/openstack/openstack-ansible/+/876160 | 22:26 |
| opendevreview | Damian DÄ…browski proposed openstack/openstack-ansible master: Deploy step-ca when 'stepca' is part of the deployment scenario. https://review.opendev.org/c/openstack/openstack-ansible/+/876637 | 22:26 |
| opendevreview | Damian DÄ…browski proposed openstack/openstack-ansible master: Add a /etc/hosts entry for the external IP of an AIO https://review.opendev.org/c/openstack/openstack-ansible/+/876638 | 22:26 |
| opendevreview | Damian DÄ…browski proposed openstack/openstack-ansible master: Use certbot to generate SSL cert for the external VIP in 'stepca' scenario https://review.opendev.org/c/openstack/openstack-ansible/+/876639 | 22:26 |
| opendevreview | Damian DÄ…browski proposed openstack/openstack-ansible master: Use a map file to select haproxy horizon backend from the base frontend https://review.opendev.org/c/openstack/openstack-ansible/+/876851 | 22:26 |
| opendevreview | Damian DÄ…browski proposed openstack/openstack-ansible master: Enable TLS frontend for repo_server by default https://review.opendev.org/c/openstack/openstack-ansible/+/876426 | 22:26 |
| opendevreview | Damian DÄ…browski proposed openstack/openstack-ansible master: Define some temporary vars for haproxy https://review.opendev.org/c/openstack/openstack-ansible/+/872328 | 22:26 |
| opendevreview | Damian DÄ…browski proposed openstack/openstack-ansible master: Prepare main repo for separated haproxy config https://review.opendev.org/c/openstack/openstack-ansible/+/871189 | 22:26 |
| opendevreview | Damian DÄ…browski proposed openstack/openstack-ansible master: Prepare main repo for separated haproxy config https://review.opendev.org/c/openstack/openstack-ansible/+/871189 | 22:36 |
| opendevreview | Damian DÄ…browski proposed openstack/openstack-ansible-haproxy_server master: Simplify haproxy_service_configs structure https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/871188 | 22:47 |
| opendevreview | Damian DÄ…browski proposed openstack/openstack-ansible-haproxy_server master: Prepare haproxy role for separated haproxy config https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/875779 | 22:47 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!